How to pickle (or handle) a generator object during tool streaming

[c0nsy]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

tool.py

for chunk in self._stream_response(response_str):
                        yield chunk
def _stream_response(self, response_str):
        """Helper function to stream response in chunks.
    
        Args:
            response_str (str): The response string to be streamed.
    
        Yields:
            str: A chunk of the response string.
        """        
        response_json = json.loads(response_str)
        answer_content = response_json['tool_output']['answer']
        chunk_size = 100  # Define your chunk size (# of characters per chunk)
    
        # Stream the answer content in chunks
        for i in range(0, len(answer_content), chunk_size):
            yield answer_content[i:i + chunk_size]
    
        # Stream the rest of the response in chunks
        response_json['tool_output']['answer'] = ""
        remaining_response_str = json.dumps(response_json)
        for i in range(0, len(remaining_response_str), chunk_size):
            yield remaining_response_str[i:i + chunk_size]

chat.py -> the error is occurring on astream_events 
```python
       try:
            events = []
            async for event in agent_executor.astream_events(agent_input, version="v1"):
                print(f"Event from astream log: {event} ",end="|", flush=True)
                events.append(event)

        except Exception as e:
            logger.error(f"Error during agent execution: {e}")
            yield "Sorry, there was an error processing your request."

### Description

My goal is to stream the response I get from a tool and then stream the final output.

I am using a custom tool that is seemingly working as intended, I am able to stream the response to my chat.py, however when I use astream_log or astream_event (not really sure which to use), when it gets to the actual llm answer portion it throws an error "Cannot pickle a generator"

I am using v1 of langchain

### System Info

root@e12630058745:/workspaces# pip freeze | grep langchain
langchain==0.1.9
langchain-aws==0.1.6
langchain-community==0.0.29
langchain-core==0.1.52
langchain-openai==0.1.1

macosx

python 3.11